{"$schema":"https://wellknown.network/schemas/agent-record-v1.json","schemaVersion":"1","id":"ag_p6qqf8jd5zd4","handle":"injection-pii-guardrail-mcp","url":"https://wellknown.network/agents/injection-pii-guardrail-mcp","links":{"self":"https://wellknown.network/agents/injection-pii-guardrail-mcp/record.json","html":"https://wellknown.network/agents/injection-pii-guardrail-mcp","markdown":"https://wellknown.network/agents/injection-pii-guardrail-mcp/record.md","api":"https://wellknown.network/api/v1/agents/injection-pii-guardrail-mcp","status":"https://wellknown.network/api/v1/agents/injection-pii-guardrail-mcp/status","claim":"https://wellknown.network/agents/injection-pii-guardrail-mcp/claim","claimApi":"https://wellknown.network/api/v1/claims","claimDescriptor":"https://wellknown.network/agents/injection-pii-guardrail-mcp/claim.json","badge":"https://wellknown.network/agents/injection-pii-guardrail-mcp/badge.svg","openapi":"https://wellknown.network/openapi.json","history":"https://wellknown.network/api/v1/agents/injection-pii-guardrail-mcp/history","tools":"https://wellknown.network/api/v1/agents/injection-pii-guardrail-mcp/tools"},"ard":{"identifier":"urn:air::server:injection-pii-guardrail-mcp","type":"application/mcp-server-card+json"},"kind":"mcp_server","declared":{"name":"injection-pii-guardrail-mcp","summary":"An MCP server providing prompt-injection detection, PII/secrets redaction, and an audit trail for AI agents.","description":"# Agent Guardrail MCP\n\nAn MCP (Model Context Protocol) server that gives any agent client — Claude\nDesktop, Claude Code, or a custom pipeline — a callable security layer:\nprompt injection detection, PII/secrets redaction, and a queryable audit\ntrail. Built to demonstrate the agent-governance primitives enterprises are\nincreasingly requiring before approving agents for production.\n\n## What it does\n\nThree concerns, exposed as four MCP tools:\n\n| Concern | Tool | What it returns |\n|---|---|---|\n| Is incoming text trying to manipulate the agent? | `scan_input(text, source)` | Risk score (0–100), risk level, matched reasons, recommendation |\n| Could outgoing text leak PII or secrets? | `scan_output(text)` | Findings list, a redacted-safe version of the text, recommendation |\n| What has the guardrail seen? | `get_audit_trail(limit, risk_level)` | Recent scan records, filterable by risk level |\n| Give me an overview | `get_guardrail_stats()` | Aggregate counts by risk level, scan type, recommendation |\n\nDetection is regex-based — no ML model, no external API call required for\nthe core path. It's fast, has zero runtime dependencies beyond the standard\nlibrary for the detectors themselves, and every decision is explainable:\nthe system tells you *which pattern matched and why*, not just a score.\n\n## Project structure\n\n```\nagent-guardrail-mcp/\n├── pyproject.toml              # packaging metadata, console script entry point\n├── requirements.txt            # for local dev without installing the package\n│\n├── guardrail/\n│   ├── __init__.py\n│   ├── server.py               # MCP server — exposes the four tools, console entry point\n│   ├── injection_detector.py   # 23 weighted regex patterns, 5 attack categories\n│   ├── pii_detector.py         # PII + credential detection and redaction\n│   └── audit.py                # append-only SQLite audit log\n│\n├── eval/\n│   ├── eval_set.json           # 35 labeled samples (22 malicious, 13 benign)\n│   └── run_eval.py             # computes pre…","publisher":{"name":"Naman","url":null},"homepage":"https://github.com/ss-2303/agent-guardrail-mcp","repository":"https://github.com/ss-2303/agent-guardrail-mcp","version":"0.1.2","license":null,"protocols":["mcp"],"tags":["guardrail","llm-security","mcp","model-context-protocol","pii","prompt-injection","security"],"pricing":null,"endpoints":[{"url":"pypi:injection-pii-guardrail-mcp","type":"package_pypi","auth":null,"probeable":false}],"skills":null,"tools":null,"extra":null,"attribution":{"kind":"pypi","name":"pypi","repoUrl":"pypi","summary":"pypi","version":"pypi","description":"pypi","homepageUrl":"pypi","publisherName":"pypi"}},"derived":{"capabilities":[{"slug":"ai.prompting","name":"Prompt Management","confidence":1,"provenance":"derived"},{"slug":"dev.package-management","name":"Packages & Dependencies","confidence":0.859,"provenance":"derived"}],"categories":["ai","dev"],"language":"en"},"observed":{"status":"unknown","statusReason":"Distributed as a package to run locally; no network endpoint to check.","lastOkAt":null,"lastProbedAt":null,"statusComputedAt":null,"reliability30d":null,"latestObservations":[],"tools":null,"package":{"name":"injection-pii-guardrail-mcp","registry":"pypi","observedAt":"2026-09-15T21:22:01.611Z","publishedAt":"2026-06-21T11:54:13.473550Z","latestVersion":"0.1.2"},"toolSurface":null,"endpointFacts":[]},"verification":{"claimed":false,"claimedAt":null,"proofs":[]},"provenance":{"sources":[{"source":"pypi","key":"injection-pii-guardrail-mcp","url":"https://pypi.org/project/injection-pii-guardrail-mcp/","firstSeenAt":"2026-09-09T22:21:41.403Z","fetchedAt":"2026-09-15T21:20:02.327Z","normalizedAt":"2026-09-15T21:20:02.327Z"}]},"firstSeenAt":"2026-09-09T22:21:41.403Z","updatedAt":"2026-09-15T21:22:01.611Z"}